CERAMICSINTERNATIONAL

Available online at www.sciencedirect.com

http://dx.doi.org/0272-8842/& 20

nCorrespondinE-mail addre

(2015) 13189–13200

Ceramics International 41 www.elsevier.com/locate/ceramint

Multiferroic (NiZn) Fe2O4–BaTiO3 composites prepared from nanopowdersby auto-combustion method

A.S. Dzunuzovica, M.M.Vijatovic Petrovica, B.S. Stojadinovicb, N.I. Ilica, J.D. Bobica,C.R. Foschinic, M.A. Zagheted, B.D. Stojanovica,n

aInstitute for Multidisciplinary Research University of Belgrade, Belgrade, SerbiabInstitute of Physics, University of Belgrade, SerbiacUNESP, Faculty for Engineering, Bauru, SP, Brazil

dUNESP, Institute for Chemistry, Araraquara, SP, Brazil

Received 24 June 2015; received in revised form 10 July 2015; accepted 16 July 2015Available online 4 August 2015

Abstract

Nickel zinc ferrite (NZF) and barium titanate (BT) were prepared by auto-combustion synthesis as an effective, simple and rapid method.Multiferroic composites with the general formula yNi1�xZnxFe2O4� (1�y)BT (x¼0.3, 0.5, 0.7, y¼0.5) were prepared by mixing NZF and BTpowders in a liquid medium in the ball mill. The FEG micrographs indicated the primary particle size less than 100 nm for both, barium titanateand nickel zinc ferrite phases. X-ray analysis and Raman spectroscopy indicated the formation of well crystallized structure of NZF and BT phasein the composite powders and ceramics, with a small contribution of the secondary phase. The homogenous phase distribution in obtainedcomposites was also confirmed. Impedance spectroscopy measurements were carried out in order to investigate the electrical resistivity ofmaterials, showing that grain boundaries have greater impact on the total resistivity than grains. Saturation magnetization and remnantmagnetization continuously decrease with barium titanate phase increase.& 2015 Elsevier Ltd and Techna Group S.r.l. All rights reserved.

Keywords: A. Powders: chemical preparation; B. Composites; C. Impedance; D. BaTiO3 and titanates; D. Ferrites

1. Introduction

Miniaturization of the solid-state electronics is achieved bydownscaling and multifunctionality. Ferroics and multiferroicsare among the most attractive multifunctional materials [1].These nanostructured materials stimulated a sharply increasinginterest to their significant technological promise in noveldevices due to fact that the combination of dissimilar materialsin ferroic-based oxide nanocomposites resulted in totally novelfunctionality.

The term multiferroic (MF) was first used by Schmid in1994. His definition referred to multiferroics as a single phasematerials which simultaneously possess two or more primary

10.1016/j.ceramint.2015.07.09615 Elsevier Ltd and Techna Group S.r.l. All rights reserved.

g author.ss: bstojanovic80@yahoo.com (B.S. Stojadinovic).

ferroic (ferroelectric, ferromagnetic and ferroelastic) properties.Today the term multiferroic has been expanded to includematerials which exhibit any type of long range magneticordering, spontaneous electric polarization, and/or ferroelasti-city. Working under this expanded definition the history ofmagnetoelectric multiferroics can be traced back to the 1960s[1–3]. In the most general sense the field of multiferroics wasborn from studies of magnetoelectric systems [4–6]. After aninitial burst of interest, research remained static until early2000. In 2003 the discovery of large ferroelectric polarizationin epitaxially grown thin films of BiFeO3 [7] and the discoveryof strong magnetic and electric coupling in orthorhombicTbMnO3 and TbMn2O5 have stimulated activity in the fieldof multiferroics. Besides scientific interest in their physicalproperties, multiferroics are interesting due to their potentialapplications as transducers, actuators, switches, magnetic field

Fig. 1. Scheme of NZF, BT and composites preparation.

Fig. 2. The X-ray diffraction patterns of (a) NZF(30–70)�BT, (b) NZF(50–50)�BT and (c) NZF(70–30)–BT powders.

Fig. 3. FT-IR spectra of (a) NF, (b) ZF, (c) BT and (d) NZF(70–30)�BTpowders.

A.S. Dzunuzovic et al. / Ceramics International 41 (2015) 13189–1320013190

sensors, new types of electronic memory devices, capacitive/inductive passive filters for telecommunications, etc. [8,9]. Alarge number of publications have been dedicated to multi-ferroics, dealing with theoretical, experimental, and applicationaspects [2,10]. In spite of hundreds of publications focused tosingle or composite multiferroic materials in the last years,they remain highly controversial concerning their preparationmethods, phase stability, intrinsic polarization and switching,ferroelectric, ferromagnetic and magnetoelectric properties,etc. [10].

Multiferroic properties can appear in a large variety ofmaterials [11]. The ferroelectric–ferromagnetic composites, astwo-phase multiferroic materials, are desired not only for thefundamental research of magneto-electric effect, but also forthe potential applications in many electronic devices [12]. Themost widely studied systems correspond to Co or Ni ferrites,with PZT, PNT, BT or BST [13]. Among them, the Ni–Znferrites/BaTiO3 systems need to be further investigatedbecause of high electrical resistivity, chemical stability andexcellent electromagnetic properties of the Ni–Zn ferrites, andhigh permittivity, low dielectric loss and high tunability ofBaTiO3 [4–6]. Those composites have attracted considerableattention as a new class of nanoferrites, expanding their use in

other areas, such as drug delivery, heterogeneous catalysis,levitated railway system, magnetic-refrigeration, microwavedevices, antennas, etc. [14].To obtain the multiferroics, several routes for conventional

material fabrication are being applied. Popular techniqueswithin the multiferroic community are: solid state synthesis,hydrothermal synthesis, sol–gel processing, vacuum baseddeposition or other wet chemical synthesis methods. However,some types of multiferroics require specific processing condi-tions within more appropriate techniques. Consequently, multi-ferroic composites request methods for the synthesis bothcomponents: ferroelectric and ferromagnetic.Ferrites crystallize in three crystal type: spinel, garnet type

and magnetoplumbite type [15,16]. Meanwhile, the mainattention is stressed to spinel type of ferrites that can besynthesized by a sol–gel method, conventional solid statereaction, mechanical attrition, hydrothermal synthesis, self-propagating combustion method, thermolysis, wet chemicalco-precipitation technique, self-propagating, microemulsion,

Fig. 4. SEM images of (a) NZF(70–30), (b) BT and (c) NZF(70–30)�BT powders.

A.S. Dzunuzovic et al. / Ceramics International 41 (2015) 13189–13200 13191

microwave synthesis, etc. [17,18]. Recently, auto-combustionsynthesis starts to be popular as rapid, cheap and rather simpletechnique. Ferroelectric component in multiferroic composites,such as barium titanate – BaTiO3 (BT) can be produced usinga huge number of various well-known methods. However, toobtain barium titanate through an advanced synthesis methodlike auto-combustion synthesis is under some difficulties dueto lack of the literature data for the preparation of the BT bythis method [5].

The aim of this study was to prepare multiferroic composites(Ni–Zn) ferrite–barium titanate from nanopowders obtained byan auto-combustion technique. It was shown that auto-combustion synthesis is very convenient for obtaining theferrite powder as a pure phase with good properties. To obtainbarium titanate by this method is not so simple due to thepossible appearance of secondary phases which later maycomplicate the process of obtaining satisfactory properties ofmultifferoic composites. A number of different methods wereused to characterize obtained powders and ceramic compositesin order to fabricate functional multiferroic material with both,ferroelectric and magnetic properties.

2. Material and methods

The multiferroic composite materials, consisting of Ni1�x

ZnxFe2O4 (x¼0.3, 0.5, 0.7, denoted as NZF(70–30), NZF(50–50), NZF(30–70) and BaTiO3 (BT), were obtained using thesynthesis route shematically presented in Fig. 1.The raw materials used for the synthesis of nickel zinc ferrite

were Fe(NO3)3 � 9H2O (Alfa Aesar, 98.0–101.0%), Ni(NO3)2 � 6H2O (Alfa Aesar, 99.9985%), Zn(NO3)2 � 6H2O (AlfaAesar, 99%), C6H8O7 �H2O (Carlo Erba, 99.5–100.5%) andNH4OH (Lach Ner, 25%). The molar ratio of Fe-ions, NiþZn-ions, citric acid was 2:1:1. Metal nitrates and citric acid solutionwere mixed by dissolving in a minimum amount of deionisedwater. The pH value of the solution was adjusted to 7 using theammonia solution. After that the solution was heated and stirred atthe temperature of about 90 1C until it converted into a xerogel,which was further heated in a heating calotte at 200 1C when self-propagation reaction was achieved. The formed powder wascalcined at 1000 1C/1 h with heating rate 2 1C/min [19].Starting reagents used for BT synthesis were Ti(OCH(CH3)2)4

(TTIP) (Alfa Aesar, 98.0–101.0%), HNO3, C6H8O7 �H2O (Carlo

Fig. 5. FEG micrographs of (a) BT, (b) NZF(70–30) and (c) NZF(70–30)–BT powders.

A.S. Dzunuzovic et al. / Ceramics International 41 (2015) 13189–1320013192

Erba, 99.5–100.5%), Ba(NO3)2 and NH4OH (Lach Ner, 25%).Firstly, ammonium hydroxide was added to TTIP solution, withconstant cooling. During this process, the yellow precipitate wasformed. It was washed with deionized water in a Buchner funnelon a vacuum pump. The obtained residue was dissolved in dilutedHNO3. Solutions of TiO(NO3)2 and BaNO3 were mixed and citricacid was added as a fuel. pH value of solution was adjusted to 6.6using NH4OH. When solution was turned to xerogel, by heating at90 1C, temperature was raised to 150 1C and self-ignition reactionoccurs. It is very fast and exothermic reaction, and gray ashformed during combustion process represents the BT precursorpowder. This powder was calcined at 900 1C for 2 h, with aheating rate of 5 1C/min (Electron-UK oven).

Multiferroic composites NZF–BT were prepared by mixingchemically obtained powders of the NZF and BT in theplanetary ball mill for 24 h. The mass ratio of NZF and BTwas always 1:1 for all obtained samples. Wolfram carbide ballsand iso-propanol were used as a milling media. The compo-sites powders were uniaxially pressed at 196 MPa into pelletsand sintered at 1200 1C for 2 h.

The phase and crystal structure analysis was carried out byX-ray diffraction technique (Rotate anode Rigaku RINT2000,

Experimental conditions: 40 kV, 60 mA. Linear detector D/teXUltra – Rigaku, Divergence slit: 0, 25, Horizontal aperture slit:5 mm). Micro-Raman spectra of the synthesized compositeswere collected at room temperature in the backscatteringconfiguration using a JobinYvon T64000 spectrometer. The514-nm laser line of a mixed Arþ /Krþ laser was used as anexcitation source with an incident laser power 60 mW in orderto minimize heating effects. The ceramic composite sampleswere measured in the range 200–800 cm�1. The FT-IR spectrawere recorded with a Bruker Equinox-55 instrument. Themorphology of the powders and microstructure of ceramicswere examined using scanning electron microscope (SEMModel TESCAN SM-300) and field emission microscope (FE-SEM, JEOL, JSM-7500F). The grain size is determined usingImageJ program. The impedance measurements were per-formed using an LCR meter (model 9593-01, HIOKI HITES-TER). Samples were prepared by coating their polishedsurfaces with Ag paste to improve the electrical contact. Thereal and imaginary parts of the complex impedance weremeasured in the frequency range of 42 Hz to 1 MHz andtemperature range of 50–200 1C and referred as the Nyquistplot. Collected data were analyzed using the commercial

Fig. 6. The XRD patterns of (a) NZF(30–70)–BT, (b) NZF(50–50)–BT and (c)NZF(70–30)–BT ceramics.

Fig. 7. (a) Raman spectra of (a) BT, (b) NZF(30–70), (c) NZF(50–50), and (d)NZF(70–30) and (b) Raman spectra of composites (a) NZF(30–70)–BT, (b)NZF(50–50)–BT and (c) NZF(70–30) at room temperature.

A.S. Dzunuzovic et al. / Ceramics International 41 (2015) 13189–13200 13193

software package Z-view. Magnetic measurements of materialswere carried out using a superconducting quantum interfero-metric magnetometer SQUID (Quantum Design).

3. Results and discussion

XRD patterns of NZF(70–30)–BT, NZF(50–50)–BT andNZF(30–70)–BT powders, presented in Fig. 2, shows thatnickel zinc ferrite, according to JCPDS files no. 10-0325, andbarium titanate phases, according to JCPDS files no. 05-0626,were obtained.

FT-IR spectra of calcined nickel ferrite (NF), zinc ferrite(ZF), BT and NZF(70–30)–BT powders, recorded in the wavenumber range of 400–4000 cm�1 at room temperatures, arepresented in Fig. 3. FT-IR spectra of NF and ZF display threebands at 560, 1640 and 3400 cm�1. Strong peak at 560 cm�1

corresponds to metal ion-oxygen complexes in the tetrahedralsites. Small peak at 1640 cm�1 was assign to the adsorbedwater or humidity. The peak at 3400 cm�1 corresponds tostretching and banding vibration of O–H bonds [20].

Absorption band near 3500 cm�1 can be observed in FTIRspectra of barium titanate. This peak was assigned to thestretching mode of internal OH� ions [21]. The broad bands at590–680 cm�1 correspond to stretching of Ti–O bond [22].

Fig. 4 shows SEM images of the pure NZF(70–30), BT andcomposite NZF(70–30)–BT powders. It is possible to noticethat the powders indicated the strong agglomeration with smallprimary particle size (o100 nm). Small particles produced bychemical synthesis usually tend to form the agglomerates, as itwas observed in the investigated case. That problem will bestudied more carefully in the future period, having in mind theimportance of composite materials properties. The use ofattrition milling after calcination could be one of the possiblesolutions for the agglomeration reduction [23].

The FE-SEM micrographs of the obtained powders (Fig. 5)mostly indicated the rounded shapes of barium titanate particleswith primary particle size less than 50 nm. The shape of ferriteparticles is pyramidal like. In the composite ferroelectric–ferro-magnetic powders obtained by homogenization of individualferroelectric and ferromagnetic phase, two separate constituentscould be clearly noticed, the one with rounded particles thatbelongs to BT and the other one with pyramidal particle shapethat belongs to ferrite, demonstrating that multifferoic compositeswere obtained with a good dispersion of the nickel–zinc ferritespinel phase in the BT ferroelectric matrix. The average particlesize in the multiferroic composite powders is below 100 nm forBT and 150–300 nm for ferrites.

The XRD difractograms for sintered samples of obtainedcomposites are presented in Fig. 6. The formation of bothphases, NZF and BT was detected (JCPDS files no. 10-0325,JCPDS files no. 05-0626). Series of small peaks at 2θ angles of32.2, 34.1, 37.1, 40.2, 54.9 and 63.11, according to JCPDSfiles no. 84-0757, indicated barium ferrite (BaFe12O19) as asecondary phase present in the sintered composites.

The relative contribution of the secondary phase, around 8%,was calculated from XRD patterns of ceramics composites,according to the most intensive peak at the 321. Therefore, it is

evident that the amount of the secondary phase is the highest inthe case of NZF(70–30)–BT. The appearance of BaFe12O19 wasnoticed in a few published articles, as well [24,25]. Therefore,from the available literature data was concluded that the presenceof small amount of this secondary phase cannot significantly affectthe magnetic properties of obtained composites. However, forfurther investigation, an effort will be done to obtain pure phasecomposites, in order to provide clearer information about influenceof secondary phases on the properties of ferroelectric–ferromag-netic composites.

Fig. 8. SEM images of (a) NZF(30–70)–BT, (b) NZF(50–50)–BT, and (c) NZF(70–30)–BT ceramics.

A.S. Dzunuzovic et al. / Ceramics International 41 (2015) 13189–1320013194

The Raman spectra of BaTiO3 and Ni1�xZnFe2O4 ceramics(x=0.3, 0.5 and 0.7) are given in Fig. 7a) for the comparison withobtained composite materials. In the spectra of NZF(70–30)ceramics the most intense Raman modes of nickel ferrite phaseare two F2g modes at 482 and 575 cm�1, one Eg mode at335 cm�1 and one A1g mode at 702 cm�1 with a shoulder at666 cm�1of Eg symmetry [26,27]. The most prominent modes ofthe zinc ferrite phase are the F2g mode at 451 cm�1 and broadA1g mode at around 647 cm�1 of low intensity [28,29]. Withincreasing content of Zn in the Ni1�xZnxFe2O4 samples, the F2g

mode of zinc ferrite becomes more intense and the Eg mode ofnickel–ferrite at 660 cm�1 shifts to the lower frequency,approaching the frequency of zinc–ferrite A1g mode. The otherF2g mode of nickel ferrite phase at 575 cm

�1 almost disappears inthe NZF(30–70) sample.The Raman spectrum of the BaTiO3 tetragonal phase

presented the most prominent Raman modes are a broad bandat about 270 cm�1 [A1 (TO)], a sharp peak at �303 cm�1 [E(TOþLO) mode], a mode at 516 cm�1[A1 (TO), E (TO)] anda mode at around 720 cm�1 [E (LO), A1 (LO)] [30].

Fig. 9. FEG microstructure of NZF(70–30)–BT composite sintered at 1200 1Cfor 2 h.

Fig. 10. Complex impedance spectra of all ceramics measured at 200 1C.

Table 1Grain resistance, grain boundary resistance, total resistance and capacitance forall samples.

Sample From Z″ to Z0 From Z″ to f

T(1C)

Rg

(Ω m)Rgb

(Ω m)Rtotal

(Ωm)Rgb

(Ω m)Cgb (nF/m)

NZF(30–70)–BT

50 155 7113 7268 4326 72.17100 117 1296 1413 822 74.45150 76 390 466 268 74.13200 22 74 96 52 72.12

NZF(50–50)–BT

50 250 54,039 54,289 30,882 109.71100 234 9231 9465 6844 135.27150 208 6562 6770 5380 160200 39 1929 1968 1504 146.24

NZF(70–30)–BT

50 585 36,480 37,065 24,590 99.63100 430 16,401 16,831 11,850 94.63150 57 492 549 398 85.87200 31 207 238 172 78.4

A.S. Dzunuzovic et al. / Ceramics International 41 (2015) 13189–13200 13195

The Raman spectra of the sintered composites are presented inFig. 7b). In the Raman spectra of NZF(70–30)–BT sample,several Raman modes of nickel ferrite suffered changes in theposition, intensity and bandwidth. The Raman mode at 482 cm�1

is asymmetrically broadened towards higher frequencies, whereasthe intensity of 575 cm�1 mode decreases. The width of the A1g

mode at 702 cm�1 increases reflecting the presence of more thantwo phases. In the fitting range from 600 to 800 cm�1 for NZF(70–30)–BT ceramics, it could be seen that this mode is well fittedwith three phases, nickel ferrite, BT phase and a secondary phase,BaFe12O19 [31,32]. This result is in agreement with previouslydiscussed XRD results.

The noticeable changes were seen in the Raman spectra ofthe NZF(50–50)–BT and NZF(30–70)–BT samples. In theNZF(50–50)–BT sample, the mode at 335 cm�1 of nickelferrite phase was substantially broadened, whereas the mode ofBT phase at around 265 cm�1 appeared. The deformation ofthe nickel ferrite mode at 482 cm�1 and its shift to higherfrequencies, due to the presence of BT phase, is obvious. Withfurther reduction of nickel ferrite phase in the NZF(30–70)–BTsample, the broad Raman mode at about 500 cm�1 iscomposed of F2g and [A1 (TO), E (TO)] modes of nickel zincferrite and BaTiO3.

SEM images of sintered ceramic samples on the free surfaceare presented in Fig. 8. Insets of the images are displayingbackscattered micrographs, demonstrating the homogenousphase distribution in obtained composites. Grains are nano-sized with different shapes, polygonal grains typical for nickelzinc ferrite, rounded grains characteristic for barium titanateand plate like grains that most likely correspond to the bariumferrite phase. All phases possess similar grain size, around1 mm. The densities of composites were 5.09 g/cm3 for NZF(70–30)–BT, 5.28 g/cm3 for NZF(50–50)–BT, 5.15 g/cm3 forNZF(30–70)–BT which corresponds to 89.6%, 93.0% and90.8% of theoretical densities, respectively, showing theincreasing trend with Zn content up to NZF(50–50)–BT andthen with further increase of Zn the density starts to decrease.Theoretical values of density are 6.01 g/cm3, 5.35 g/cm3,5.36 g/cm3 and 5.33 g/cm3 for the pure BT, NZF(70–30),NZF(50–50) and NZF(30–70) phases, respectively. Achievedceramics densities are rather small due to high agglomerationand this can evidently affect the electrical (dielectric, ferro-electric) properties. This problem can be possibly solved by atreatment in the attrition mill, which may enable the prepara-tion of ceramic composites with improved properties. The FE-SEM microstructure of composites is presented in Fig. 9. It ispossible to notice two different phases: ferrites (platelike orpyramidal grains) and barium titanate with more roundedgrains. It is important to notice that no reaction betweenparent components was observed.

The impedance spectroscopy (IS) is important method to studythe electrical properties of one material, because it gives theinformation about resistive and reactive components in thematerial. IS was used to evaluate the contributions of variouscomponents such as grain and grain boundary to the overallelectrical properties of NZF–BT ceramics composites. Fig. 10shows the impedance plots for NZF(70–30)–BT, NZF(50–50)–BT,

A.S. Dzunuzovic et al. / Ceramics International 41 (2015) 13189–1320013196

NZF(30–70)–BT ceramics at 200 1C. The impedance spectra wereanalyzed using commercially available Z-View software. For allsamples, a one depressed semicircular arcs is present, indicating thepossible overlapping of two arcs that correspond to grain and grainboundary contributions. The Z view software and equivalent circuitconsisted of two parallel R-CPE elements connected in series,which were used to evaluate the grain boundary resistivity at lowfrequencies and grain contribution at high frequencies. Theexistence of two different phases (ferroelectric and ferromagnetic)in one composite material can make the interpretation of impe-dance results of these materials rather complicated. The appearanceof two semicircular arcs could be also the indication of thepresence of two different crystallographic phases. When comparedto resistivity values of pure BT and NZF phases [19,33], quite adifference in the resistivity magnitude can be observed. BT,ferroelectric phase, is more resistive in comparison with NZFphase, indicating its dominant effect in the total resistivity of thecomposite materials. The values of the grain, grain boundary andtotal electrical resistivity of obtained ceramic composites arepresented in Table 1.

With increasing temperature, the resistance of the grain andthe grain boundary decreases for all composites (Table 1), as itwas expected. Comparing the total resistance at the same

Fig. 11. Arrhenius plots of grain (σg) and grain boun

temperature, it can be noticed that NZF(50–50)–BT possessesthe highest values. With increase of Zn content up to 50% theresistance increases, probably because zinc leads to betterstructure ordering and results in the reduction of defects. Mostlikely, oxide ion vacancies Vo�� were formed due to loss ofoxygen in the sintering process and present the main con-ductive species in the obtained ceramics [19].The temperature dependence of the resistance can be

presented by the equation [34]:

σ ¼ σ0exp � Ea

kbT

� �ð1Þ

where Ea, σ0 and kb are the activation energy of the carriers forconduction, the pre-exponential factor and the Boltzmann constant,respectively. Arrhenius plots of the grain, σg, and grain boundaryconductivity, σgb, for NZF(30–70)–BT, NZF(50–50)–BT and NZF(70–30)–BT ceramics are presented in Fig. 11. The activationenergy can be calculated from the slope of the given diagrams. Thevalues of the grain and grain boundary activation energies were:Ea(g)¼0.15 eV, Ea(gb)¼0.39 eV for NZF(30–70)–BT, Ea(g)¼0.17 eV, Ea(gb)¼0.27 eV for NZF(50–50)–BT and Ea(g)¼0.28 eV, Ea(gb)¼0.49 eV for NZF(70–30)–BT. Generally, theactivation energies for electron hopping are lower than for hole

dary (σgb) conductivity for all ceramic samples.

A.S. Dzunuzovic et al. / Ceramics International 41 (2015) 13189–13200 13197

hopping. In the polaron conduction, the activation energy of holesis usually used to be greater than 0.2 eV [35]. On this basis, thecalculated activation energies indicate that the conduction inobtained materials is a consequence of polaron hopping. Theconduction in ferrites, ferroelectrics and its composites can beexplained by polaron hopping process among the localized sites.Hopping conduction is favored in ionic lattices in which the samekind of cation is found in two different states [36]. Thus, in theobtained composite materials the hopping of 3d electrons amongFe2þ and Fe3þ as well as between Ni2þ and Ni3þ in the ferritephase and also Ti4þ to Ti3þ in the ferroelectric phase, could playan important role in the conduction processes. These values for Eaare also in agreement with the results obtained for bariumstrontium titanate–nickel zinc ferrite composites in the study ofother authors [25,37]. The activation energy for the conductionprocess through the grain boundaries in all measured samples washigher compared to the values of the activation energy for theconduction process through the grains. Therefore, the grainboundary effect can be ascribed as the dominant effect in totalconduction of ceramic composites.

Complex modulus plots (M″–M0) at the different tempera-tures for NZF(30–70)–BT, NZF(50–50)–BT, NZF(70–30)–BT

Fig. 12. Modulus complex plots of (a) NZF(30–70)–BT, (b) NZF(5

are shown in Fig. 12. Values of M″ and M0 were calculatedfrom equations M″¼ωCoZ0 and M0 ¼ωCoZ″, respectively,where Co is equal to εoA/h and angular frequency ω is equalto 2πf. The complex electric modulus analysis can be used toseparate the electrode polarization effect from the grainboundary conduction process and also to determine theconductivity relaxation times [25]. In the presented diagramstwo semicircular arcs can be noticed in NZF(70–30)–BT, NZF(50–50)–BT and NZF(30–70)–BT composites, even though,they were not fully resolved in the complex impedance plots.This is also the indication of the presence of differentrelaxation processes, due to contribution of grain, grainboundary and/or different crystallographic phases.Complex impedance plots Z″–f for all ceramic composites

are presented in Fig. 13. With increasing temperature thepositions of peaks shift toward high frequency side, whichleads to the conclusion that the dielectric relaxation isthermally activated process [38]. A relative lowering in themagnitude of Z″ with a shift of the peaks towards the higherfrequency arises is possibly due to the presence of spacecharge polarization or accumulation at the grain boundaries[25]. Broadening of the peaks with rise in temperature can be

0–50)–BT, and (c) NZF(70–30)–BT at different temperatures.

Fig. 13. Frequency dependence of imaginary parts of the impedance spectra (Z″) of (a) NZF(30–70)–BT, (b) NZF(50–50)–BT, and (c) NZF(70–30)–BT at differenttemperatures.

Fig. 14. Magnetic hysteresis loops at room temperature for all compositesceramics.

Table 2Saturation magnetization moment, saturation fields, residual magnetization andcoercive field for yNi1�xZnxFe2O4� (1�y)BT (x¼0.3, 0.5, 0.7, y¼0.5)composite materials.

Sample Msat (emu/g) Hsat (kOe) Mr (emu/g) Hc (Oe)

NZF (70–30)–BT 23.64 2.55 4.31 19.8NZF (50–50)–BT 22.23 2.05 2.89 25.4NZF (30–70)–BT 10.42 1.95 0.17 17.6

A.S. Dzunuzovic et al. / Ceramics International 41 (2015) 13189–1320013198

an indication of the presence of immobile species at lowtemperature and defects at higher temperatures [39]. Thecorresponding capacitances and resistivities were calculatedfrom the maximum value of the Z″ peak and associated fmax

value, using the relationships C¼1/2πfmaxR, where R is equalto 2Z″max [40]. The obtained values for C and R are presentedin Table 1. Obtained resistivity data were used for thecalculation of activation energies of relaxation processes at

the grain boundary and they are in accordance with Ea

obtained from the complex impedance analysis.Magnetization results are presented in Fig. 14 and Table 2.

In pure nickel zinc ferrite, with addition of Zn, magnetizationincreases because the balance between Fe3þ ions in tetrahedraland octahedral sites, which is conditioned by migration of theFe3þ ions inter this sites. This leads to the weakening of A–Bexchange interaction, causing the increase of magnetization.The change of the magnetic properties with reducing amountof magnetic phase was expected. In composites, a dilutioneffect exists, which means that BT does cause intimate changein the magnetic properties, leading to the reduction of themagnetic moment [41]. The saturation magnetization and

A.S. Dzunuzovic et al. / Ceramics International 41 (2015) 13189–13200 13199

remnant magnetization have shown a decrease with increasingZn content, likely due to already mentioned dilution effect. Ms,at first slightly decreases with increase of Zn content up to0.5 mol% and then significantly decreases with further increaseof Zn. In comparison with pure NZF phases, Ms also decreases,due to the presence of non-magnetic barium titanate phase[19]. The fields at which saturation occurs were around 2 kOefor NZF(30–70)–BT and NZF(50–50)–BT and slightly higherfor NZF(70–30)–BT. The coercive field, the field required toovercome the defects in the material, was higher for compo-sites in the comparison with pure NZF phases, which can beexplained by the fact that the composite possesses a higheranisotropy field than the NZF at the same applied field [35].The coercive field was the highest for NZF(50–50)–BTceramic composites.

4. Conclusion

A series of nickel zinc ferrite–barium titanate with generalformula yNi1�xZnxFe2O4� (1�y)BT (x¼0.3, 0.5, 0.7, y¼0.5)were successfully prepared by mixing of previously preparedpowders of nickel zinc ferrite and barium titanate. The compositesmaterials formation was confirmed by XRD and Raman spectro-scopy for powders and ceramics, with small amount of secondaryphase. SEM analysis indicated a high level of powder agglomera-tion and ceramic composites with different shapes grains of around1 μm. Impedance analysis has shown that for all obtained ceramicsamples grain boundary resistance has the highest contribution tothe total resistance. The resistance was found to decrease with theincreasing temperature for all measured samples, and it can benoticed that NZF(50–50)–BT possesses the highest values of totalresistance. The calculated activation energy indicates that theconduction in the composite materials is a consequence of polaronhopping. Results of the magnetic measurements show that themagnetization in the composites ceramics is reduced in comparisonwith pure nickel zinc ferrite ceramics due to the presence of bariumtitanate phase. The fields at which saturation occurs were almostthe same for all investigated compositions.

Acknowledgments

The authors gratefully acknowledge the financial support ofthe Ministry of Education, Science and Technological Devel-opment of the Republic of Serbia (Project III 45021), COSTMP 0904 and IC 1208. Special thanks to Dr. LaviniaCurecheriu from Faculty of Physics, Al.I. Cuza University,Iasi, Romania for magnetic measurements and to Dr. SoniaZanetti from Institute for Chemistry UNESP, Araraquara,Brazil for FEG microstructure measurements. Also we wouldlike to thank Dr. Zorana Dohcevic-Mitrovic from Institute ofPhysics, University of Belgrade for useful suggestions aboutRaman spectroscopy measurements.

References

[1] H. Schmid, Multi-ferroic magnetoelectrics, Ferroelectrics 162 (1994)317–338.

[2] T. Kimura, T. Goto, H. Shintani, K. Ishizaka, T. Arima, Y. Tokura,Magnetic control of ferroelectric polarization, Nature 426 (2003) 55–58.

[3] N. Hur, S. Park, P.A. Sharma, J.S. Ahn, S. Guha, S.W. Cheong, Electricpolarization reversal and memory in a multiferroic material induced bymagnetic fields, Nature 429 (2004) 392–395.

[4] L. Mitoseriu, Magnetoelectric phenomena in single-phase and compositesystems, Bol. Soc. Esp. Ceram. 44 (3) (2005) 177–184.

[5] M.E. Botello-Zubiate, D. Bueno-Baques, J. de Frutos, L.E. Fuentes,J.A. Matutes-Aquino, Synthesis and magnetoelectric characterization ofcobalt ferrite–barium titanate composites using a new pulsed magneticfield method, Integr. Ferroelectr. 83 (2006) 33–40.

[6] C.W. Nan, M.I. Bichurin, S. Dong, D. Viehland, G. Srinivasan, Multi-ferroic magnetoelectric composites: historical perspective status andfuture directions, J. Appl. Phys. 103 (2008) 031101.

[7] M.A. Khan, T.P. Comyn, A.J. Bell, Growth and characterization of tetragonalbismuth ferrite–lead titanate thin films, Acta Mater. 56 (2008) 2110–2118.

[8] N. Ikeda, H. Ohsumi, K. Ohwada, K. Ishii, T. Inami, K. Kakurai,Y. Murakami, K. Yoshii, S. Mori, Y. Horibe, H. Kito, Ferroelectricityfrom iron valance ordering in the charge-frustrated system LuFe2O4,Nature 436 (2005) 1136–1138.

[9] S.W. Cheong, M. Mostovoy, Multiferroics: a magnetic twist forferroelectricity, Nat. Mater. 6 (2007) 13–20.

[10] W. Eerenstein, N.D. Mathur, J.F. Scott, Multiferroic and magnetoelectricmaterials, Nature 442 (2006) 759.

[11] D.B. Litvin, Ferroic classifications extended to ferrotoroidic crystals, ActaCryst. A 64 (2008) 316–320.

[12] T. Lottermoser, T. Lonkai, U Amann, D. Hohlwein, J. Ihringer, M. Fiebig,Magnetic phase control by an electric field, Nature 430 (2004) 541–544.

[13] B.B. van Aken, T.T. Palstra, A. Filippetti, N.A. Spaldin, The origin offerroelectricity in magnetoelectric YMnO3, Nat. Mater. 3 (2004) 164–170.

[14] H. Kavas, N. Kasapoglu, A. Baykal, Y. Koseoglu, Characterization ofNiFe2O4 nanoparticles synthesized by various methods, Chem. Pap. 63(4) (2009) 450–455.

[15] J.V. Boomgaard, R.A.J. Born, A sintered magnetoelectric compositematerial BaTiO3–Ni(Co, Mn)Fe2O4, J. Mater. Sci. 13 (1978) 1538–1548.

[16] N.R. Reddy, E. Rajagopal, K.V. Sivakumar, K.K. Patankar, V.R.K. Murthy,Effect of temperature on the elastic and anelastic behaviour of magneto-ferroelectric composites Ba0.8Pb0.2TiO3þNi0.93Co0.02Mn0.05Fe1.95O4�δ in theferroelectric rich region, J. Electroceram. 11 (2003) 167–172.

[17] G.R. Amiri, M.H. Yousefi, M.R. Abolhassani, S. Manouchehri,M.H. Keshavarz, S. Fatahian, Magnetic properties and microwave absorptionin Ni–Zn and Mn–Zn ferrite nanoparticles synthesized by low-temperaturesolid-state reaction, J. Magn. Magn. Mater. 323 (2011) 730–734.

[18] D. Chen, R. He, Synthesis of nickel ferrite nanoparticles by sol–gelmethod, Mater. Res. Bull. 36 (2001) 1369–1377.

[19] A.S. Džunuzović, N.I. Ilić, M.M. Vijatović Petrović, J.D. Bobić,B. Stojadinović, Z. Dohčević-Mitrović, B.D. Stojanović, Structure andproperties of Ni–Zn ferrite obtained by auto-combustion method,J. Magn. Magn. Mater. 374 (2015) 245–251.

[20] S. Zahi, Synthesis, permeability and microstructure of the optimal nickel–zinc ferrites by sol–gel route, J. Elctromagn. Anal. Appl. 2 (2010) 56–62.

[21] M.H. Frey, D.A. Payne, Grain-size effect on structure and phasetransformations for barium titanate, Phys. Rev. B 54 (1996) 3158–3168.

[22] P.R. Arya, P. Jha, G.N. Subbanna, A.K. Ganguli, Polymeric citrateprecursor route to the synthesis of nano-sized barium lead titanates,Mater. Res. Bull. 38 (2003) 617–628.

[23] M.M. Vijatovic Petrovic, J.D. Bobic, A.M. Radojkovic, J. Banys,B.D. Stojanovic, Improvement of barium titanate properties induced byattrition milling, Ceram. Int. 38 (2012) 5347–5354.

[24] L. Mitoseriu, V. Buscaglia, M. Viviani, M.T. Buscaglia, I. Pallecchi,C. Harnagea, A. Testino, V. Trefiletti, P. Nanni, A.S. Siri, BaTiO3–

(Ni0.5Zn0.5)Fe2O4 ceramic composites with ferroelectric and magneticproperties, J. Eur. Ceram. Soc. 27 (2007) 4379–4382.

[25] K. Verma, S. Sharma, Impedance spectroscopy and dielectric behavior inbarium strontium titanate-nickel zinc ferrite composites, Phys. StatusSolidi B 249 (2012) 209–216.

A.S. Dzunuzovic et al. / Ceramics International 41 (2015) 13189–1320013200

[26] G. Dixit, J.P. Singh, R.C. Sriavastava, H.M. Agrawal, Study of 200 MeVAg15þ ion induced amorphisation in nickel ferrite thin films, Nucl.Instrum. Methods B 269 (2011) 133-139.

[27] A. Ahlawat, V.G. Sathe, V.R. Reddy, A. Gupta, Mossbauer, raman andX-ray diffraction studies of superparamagnetic NiFe2O4 nanoparticlesprepared by sol–gel auto-combustion method, J. Magn. Magn. Mater. 323(2011) 2049–2054.

[28] A. Ahlawat, V.G. Sathe, Raman study of NiFe2O4 nanoparticles, bulk andfilms: effect of laser power, J. Raman Spectrosc. 42 (2011) 1087–1094.

[29] J.P. Singh, G. Dixit, R.C. Srivastava, H.M. Agrawal, Ravi Kumar,Formation of latent tracks and their effect on the structural and magneticproperties of nanosized zinc ferrite, J. Alloy. Compd. 551 (2013)370–375.

[30] Y. Shiratori, C. Pithan, J. Dornseiffer, R. Waser, Raman scattering studieson nanocrystalline BaTiO3, J. Raman Spectrosc. 38 (2007) 1288–1299.

[31] X. Liu, J. Wang, J. Ding, M.S. Chen, Z.X. Shen, The effects ofmechanical activation in synthesizing ultrafine barium ferrite powdersfrom co-precipitated precursors, J. Mater. Chem. 10 (2000) 1745–1749.

[32] J. Kreisel, S. Pignard, H. Vincent, J.P. Senateur, G. Lucazeau, Ramanstudy of BaFe12O19 thin films, Appl. Phys. Lett. 73 (1998) 1194.

[33] S. Sharma, K. Shamim, A. Ranjan, R. Rai, P. Kumari, S. Sinha,Impedance and modulus spectroscopy characterization of lead freebarium titanate ferroelectric ceramics, Ceram. Int. 41 (2015) 7713–7722.

[34] M. Shah, M. Nadeem, M. Idrees, M. Atif, M.J. Akhtar, Change ofconduction mechanism in the impedance of grain boundaries inPr0.4Ca0.6MnO3, J. Magn. Magn. Mater. 332 (2013) 61–66.

[35] M. Atif, M. Nadeem, Interplay between the ferromagnetic and ferro-

electric phases on the magnetic and impendance analysis of (x)Pb

(Zr0.52Ti0.48)O3� (1�x)CoFe2O4 composites, J. Alloy. Compd. 623

(2015) 447–453.[36] C.E. Ciomaga, A.M. Neagu, M.V. Pop, M Airimioaei, S. Tascu,

G. Schileo, C. Galassi, L. Mitoseriu, Feroelectric and dielectric properties

of ferrite–ferroelectric ceramic composites, J. Appl. Phys. 113 (2013)

074103.[37] M.I. Klinger, Two-phase polaron model of conduction in magnetite-like

solids, J. Phys. C: Solid State Phys. 8 (1975) 3595.[38] F Borsa, D.R. Torgeson, S.W Martin, H.K. Patel, Relaxation and

fluctuations in glassy fast-ion conductors: wide-frequency-range NMR

and conductivity measurements, Phys. Rev. B 46 (1992) 795.[39] A.K. Behera, N.K. Mohanty, B. Behera, P. Nayak, Impendance properties

of 0.7(BiFeO3)–0.3(PbTiO3) composite, Adv. Mater. Lett. 4 (2) (2013)

141–145.[40] F.D. Morrison, D.C. Sinclair, A.R. West, Characterization of lanthanum

doped barium titanate ceramics using impedance spectroscopy, J. Am.

Ceram. Soc. 84 (3) (2001) 531–538.[41] L.P. Curecheriu, M.T. Buscaglia, V. Buscaglia, L. Mitoseriu,

P. Postolache, A. Ianculescu, P. Nanni, Functional properties of

BaTiO3–Ni0.5Zn0.5Fe2O4 magnetoelectric ceramics prepared from pow-

ders with core–shell structure, J. Appl. Phys. 107 (2010) 104–106.